@jmoiron.bsky.social replied: I've been choosing the Butterfly and spreading the religion of Bluesky.
jmoiron plays the blues
@jmoiron.bsky.social This cred.blue/jmoiron.bsky... site called me "barely active" and says my style of posting "unknown" and I think I'm offended.
100% alt text rate at least.
@jmoiron.bsky.social From 0 to running a local LLM model to answer the most basic coding question of them all, from the perspective of someone old, out of touch, but curious:
jmoiron.net/blog/contact...
@jmoiron.bsky.social Civ VII: Achingly beautiful. Love leader/faction split, commanders, no builders. Need some more time with ages. Narration is gorgeous. Main theme good, music great.
Baffled how new civs are always missing at least one type of research/production queue. Minimalist UI a bad fit, I...
@jmoiron.bsky.social replied: I think this is a Bad Approach, but there are limited hands and much work, and there are always enough users who have gone through the rites and hold the sacred knowledge of what "bwrap: setting up uid map: Permission denied" means who think this situation is fine and ignorance i...
@jmoiron.bsky.social replied: Ubuntu has gone through effort to provide AA profiles for a lot of packages that require these namespaces, but they only configure a specific install path, so they're pretty brittle. The approach to limiting user interruption is to encourage packagers to ship AppArmor profiles in...
@jmoiron.bsky.social replied: If you did not guess 3, then congratulations, you have never experienced Linux on the Desktop.
To fix, your options are:
1. Manually create an AppArmor profile for it, which is a ~7 line file with custom syntax
2. Disable this security feature globally through sysctl
@jmoiron.bsky.social For security purposes, Ubuntu disabled unprivileged user namespaces in 23.10. If you try to run an unconfigured app that requires them, what do you think happens?
1. An escalation prompt to configure the app
2. A popup explaining how to fix it
3. Mysterious technical error mess...
@jmoiron.bsky.social The 25 year old Nazi working for doge is a "kid", but a black child killed by police is a "male."
@jmoiron.bsky.social replied: In the wake of the Challenger disaster, Feynman told the committee that "reality must take precedence [...] because nature can't be fooled", but the pandemic snapped something in my belief that reality might ever intervene on these political fantasies.
@jmoiron.bsky.social replied: I immediately thought of this sketch when I saw this in the news. Immediately.
@jmoiron.bsky.social replied: I've written about this in the past; 13 years ago, in fact. I made some pretty shaky predictions here, but the final one looks pretty good to me.
jmoiron.net/blog/the-too...
@jmoiron.bsky.social replied: To get a better feeling for how this works, consider the first time you saw an algorithm like quicksort. Explaining what its parts do is pretty simple, but understanding why it works is more difficult. That's where the complexity went.
Functional programming is a goldmine for th...
@jmoiron.bsky.social replied: A formative insight on this subject for me is from Chuck Moore via Elizabeth Rather in this usenet post. A simple solution can solve complex problems, but that complexity hasn't gone away; it's embedded in the sophistication of the design of the solution.
groups.google.com/g/com...
@jmoiron.bsky.social replied: Someone lacking context for a problem space will find difficulty even with low complexity, whereas someone who has a lot of experience might have mental shorthands that make it much easier to navigate. The complexity is the same for both people, the difference is their experience...
@jmoiron.bsky.social replied: Complex is not merely a synonym for "difficult" or "confusing." It is a system property; it implies an intricacy, a large number of parts and/or interactions between those parts.
@jmoiron.bsky.social The core message of this post is +100, but I want to call out its usage of the word "complex", because I think it's a very useful concept for discussing software, but not the way it's used here.
@jmoiron.bsky.social In light of Armin's post, I'll bump something I wrote about the high cost of dependencies 9 years ago. I still strongly believe that avoiding deps where possible has better long term survival characteristics.
jmoiron.net/blog/depende...
@jmoiron.bsky.social Cargo was written by ruby/js devs and iirc Katz was pretty open at the time that an npm/gem like experience was the gold standard and should be the goal.
It's no surprise that this is the result, I think it's a natural consequence of some aspects of that approach.
@jmoiron.bsky.social With apologies to Tony Bennet, it appears I'll be Leaving My Gallbladder in Hong Kong.
Good riddance.